Import Your Data

Upon reaching the knowledge base creation page, the name will be “Untitled Source”. The user can change the name by clicking the pencil icon and entering the new name. There are also various choices of data sources available for the user to build the knowledge base.

Figure: Import Data Interface and Highlighting to edit the Knowledge Base Name

NOTE: Knowledge Base can create by uploading Documents (txt, pdf, docx, csv, xlsx, pptx, json, and md), by uploading Scanned Documents (pptx, pdf, docx), by crawling Websites (Single Link, and Crawl URLs), by uploading Media (jpg, jpeg, png, mp4, mp3, wav, and ogg), by uploading Files from Cloud Drives (Google Drive, Amazon S3 Buckets, OneDrive, and SharePoint).

2.2.1. Import data with Documents

Step 1: The user can import the data through Documents by simply clicking the “Browse Files” button. Then select the file with the available file type to be uploaded. It will take some time to upload the file according to its size.

Figure: Uploaded PDF File from the Documents Section for the Knowledge Base

After the file is successfully uploaded, the system will determine the complexity level of it.

Figure: Uploaded PDF File with the Complexity Status in the Knowledge Base

Step 2: If the complexity level is medium or high, the user is advised to toggle on the Advanced Document Processing slider for better processing operation.

Figure: Highlighting the “Advanced Document Processing” Toggle Slider

Step 3: After toggling on the slider, the OCR Language column will appear beside the complexity column. The user can choose the language from the file in drop-down list for better operation within the system.

Figure: Highlighting the OCR Language Column and drop-down list

Step 4: The user can hover on the file to see the “View” and “Delete” button. The user can view or delete the file by using those buttons.

2.2.2. Import data with Scanned Documents

Step 1: The user can import the data through Documents by simply clicking the “Browse Files” button. Then select the file to be uploaded.

Figure: Click the “Browse Files” button to select the File from the User’s Device

Figure: Uploaded PDF File from the Scanned Documents Section for the Knowledge Base

Step 2: After the file is successfully uploaded, the user can toggle on Document Processing slider to run the operation better for medium and high complexity files.

Step 3: After toggling on the slider, the user can choose the model for processing and editing instructions for the engine.

Figure: Highlighting to choose the Document Processing Model and Edit the Instructions Button

Step 4: The user can choose a suitable model for document processing from the drop-down list for their personalized documents.

Figure: Highlighting Processing Model Drop-down List

Step 5: The user can click “Edit Instructions” to write the instructions for the advanced processing.

Figure: Document Processing Instructions Pop-up Window for clicking “Edit Instructions”

NOTE: The user can also choose the pre-built templates by clicking the “Use templates” button.

Figure: Highlighting “Use templates” for using the Pre-built Templates

Figure: Highlighting the Pre-built Templates and the Preview Instructions in the Select a Template Pop-up Window

The user can hover on the file to see the “Delete” button.

Figure: Highlighting the “Delete” Button to delete the Uploaded Scanned Document File

2.2.3. Import data with Websites

Step 1: The user can import the data from the website by simply entering the website link in the text box. There are two ways that the users can take the data from the website: Single Link method and the Crawl URLs method.

NOTE: Single link will scrap the one exact URL provided by the user.

Crawl URLs will start from a seed URL and follow internal links recursively.

Figure: Highlighting “Single Link and Crawl URLs” in the Websites Section

Step 2: Choosing method and entering link.

Option 1: Single Link

Enter the website URL in the Website link text box. Then, click the “Add link” button to include that exact URL in the training list.

Figure: Highlighting “Website Link” Textbox and “Add Link” button

Figure: The Sample Single Link and “Add Link” Button

Figure: The Added Single Link in the URLs List

Option 2: Crawl URLs

Enter the website URL in the Website link text box. Then, click “Crawl links” button to include the link and its internal links in the training list.

Figure: Enter the link and click “Crawl links” button

After taking a while, the original URL and crawled URLs will appear in a new window. The user can use the trash icon to remove unwanted links or click the “Confirm Links” button to include the links in the training list.

Figure: The Crawled URLs in the URLs List

The user can click "Confirm Links" to confirm the crawled website links if the crawled links will not be removed from the list.

Figure: The crawled URLs in a New Pop-up Window and Highlighting “Delete Icon” and “Confirm Links” Button

Step 3: Toggle Image Detection

The user can toggle on Image Detection to extract text and content from image tags.

Step 4: After toggling on the slider, the user can choose the model for the image detection from the website and edit instructions for the engine.

Figure: Highlighting to select the Model from the Drop-down Button and Edit Instructions Button

Step 5: The user can choose a suitable model for the image detection from the drop-down list for their data.

Figure: Highlighting “Engine drop-downlist"

Step 6: The user can click “Edit Instructions” to write the instructions for the engine.

Figure: Image Detection Instructions Pop-up Window

Step 7: The user can hover on the URL and click the trash icon to delete the URL from the URLs list.

Figure: Highlighting the Delete Icon to delete the URL

2.2.4. Import data with Media

Step 1: The user can import the data through Media by simply clicking the “Browse Files” button. Then select the file to be uploaded.

Figure: Click the “Browse Files” button to select the File from the User’s Device

Figure: The Uploaded PNG File in the Media Section

Step 2: Toggle Media Processing

The user can enable the Media Processing toggle to configure the OCR/view model and instructions for images.

Step 3: After toggling on the slider, the user can choose the engine for processing and editing instructions for the engine.

Figure: Highlighting the “Engine drop-down and Edit instructions button”

Step 4: The user can choose a suitable model from the drop-down list for their data.

Step 5: The user can click “Edit Instructions” to write the instructions for the engine.

Figure: Image Detection Instructions window after clicking “Edit Instructions”

The user can also choose the pre-built templates by clicking the “Use templates” button.

Figure: Highlighting “Use templates” button

Figure: Highlighting “Pre-built templates and the instruction of the template” in Select a Template window

Step 6: The user can click the “trash can” button to delete the added media.

Figure: Highlighting the “trash can” icon

2.2.5. Import Data with Cloud Drives

Step 1: The user can import data from cloud services. Currently, the system supports four cloud services, Google Drive, Amazon S3, OneDrive, and SharePoint.

Figure: Cloud Services the system supports

Option 2: Amazon S3

Click the “Import Files” button under the Amazon S3 option.

Figure: Highlighting the “Import Files” button of Amazon S3

Sub-step 2: The window to enter credentials will appear. After entering the credentials, click the “Connect” button.

Figure: Highlighting the text box to enter credentials

Figure: Click “Connect” button after entering credentials

Sub-step 3: A list of files will appear. The user can select the files they want to upload and click the “Select” button.

Figure: A list of files and highlighting the “Select” button

Sub-step 4: The file is successfully uploaded, and the complexity is determined by the system.

Figure: File is uploaded to the system and Complexity is determined